"The modern global economy runs on identity. Every transaction, account, organization, and asset depends on knowing who is who. When that knowledge breaks down, the consequences scale quickly, especially in fraud, financial crime, and national security. Roughly $3 trillion in illicit funds moves through the global financial system each year, flowing through shell companies, false identities, and fragmented records designed to obscure accountability."
"At ODSC AI West 2025, Paco Nathan explored how entity resolution and investigative graph systems form the backbone of downstream AI applications in anti-fraud, intelligence, and risk analysis. The deeper insight running through the session was not just technical, but philosophical: AI systems are most effective when they reduce inhuman complexity back to human scale Why Fraud Is a Graph Problem, Not a Checklist Financial crime doesn't behave like a rules-based system. Criminal networks exploit relationships between people, companies, jurisdictions, and transactions, making linear detection methods ineffective."
"As John Lambert of Microsoft Threat Intelligence famously observed, defenders think in checklists, while attackers think in graphs. That asymmetry explains why fraud continues to rise even as other crime categories decline. In the UK, fraud now accounts for over 40% of all crime, despite falling rates of violent offenses. The implication is clear: stopping modern fraud requires tools that reason across networks of relationships, not isolated events."
The modern global economy depends on reliable identity across transactions, accounts, organizations, and assets. Failures in identity scale quickly in fraud, financial crime, and national security, with roughly $3 trillion in illicit funds moving annually through shell companies, false identities, and fragmented records. Entity resolution and investigative graph systems enable AI applications in anti-fraud, intelligence, and risk analysis by reasoning across relationships. Modern financial crime exploits networks, so graph-based tools outperform checklist rules. Attackers leverage relational structures across people, companies, jurisdictions, and transactions. The Azerbaijani Laundromat exemplified this, moving large sums through almost 500 shell companies and generating 17,000 leaked records used as ground truth.
Read at Medium
Unable to calculate read time
Collection
[
|
...
]